Skip to content

Latest commit

 

History

History
 
 

counterfactual_fairness

README.md

Path-Specific Counterfactual Fairness (AAAI 2019)

Paper: Path-Specific Counterfactual Fairness

If you use the code here please cite this paper:

@inproceedings{chiappa2019path,
  title={Path-specific counterfactual fairness},
  author={Chiappa, Silvia},
  booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
  volume={33},
  number={01},
  pages={7801--7808},
  year={2019}
}

Overview

This release contains the path-specific counterfactual fairness method used in the paper, as well as utility functions for loading the Adult dataset.

The following gives a brief overview of the contents, more detailed documentation is available within each file:

  • causal_network.py: Defines the Node class, instances of which can be combined into a directed graph. Associated with each node is a distribution_module, a haiku module which builds a tensorflow Distribution instance as a function of the node's parents.
  • util.py: Miscellaneous utility functions.
  • variational.py: Class for performing variational inference. The Variational haiku module is a general-purpose approximate posterior, using an MLP to map from arbitrary inputs to the parameters of a Gaussian distribution.
  • adult.py: Utility functions for the Adult dataset.
  • adult_pscf.py: Training and 'fair' prediction process (using path-specific counterfactual fairness) on the Adult dataset.
  • adult_pscf_config.py: Configuration file with default training parameters.

Dataset

The Adult dataset can be downloaded from https://archive.ics.uci.edu/ml/datasets/adult. You may use the following command to download both necessary files to run the training script:

sh download_dataset.sh ${OUTPUT_DIR}

Experiments

To download the dataset and run the main experiment reported in the paper, you may run:

sh run_adult_pscf.sh

Acknowledgements

Credits to Thomas P. S. Gillam for the original TF1 implementation.